top button
Flag Notify
    Connect to us
      Site Registration

Site Registration

Why we need SGW/PGW separately in LTE?

+5 votes

Why we need SGW and PGW separately?
If it is for roaming then how it will be used at roaming condition?

posted Nov 24, 2013 by sivanraj

Share this question
Facebook Share Button Twitter Share Button LinkedIn Share Button

4 Answers

+4 votes

Its not about LTE rather general Networking question.

Telecom network is a Hierarchical network, technically you can have all logical node functionality in a single box i.e. eNodeB. But assume that all eNodeB have big capability then how much will be the cost as eNodeB are many (in millions) and close to the UE. Also the level of connectivity is required to connect them O(n^2) where n is the number of nodes.

So to reduce the cost we break the functionality and move the functionality step by step away from UE in this way we reduce the cost of implementation drastically and ultimately the call rate. And while the designing any telecom system it is one of the prime factor (along with other i.e. backward compatibility, data-rate, cabling etc).

answer Nov 25, 2013 by Salil Agrawal
+3 votes

Both are the logical entities. It is not necessary that both should be in different hardware. There is a product called "EPC in a box" . In which all the nodes resides in same hardware.

With request your question in roaming scenario, UE sends attach request to visited network's MME and this MME talks to home HSS of UE to get subscription profile. In roaming scenarios, UE will use visited network SGW but pgw of its home network.

answer Nov 25, 2013 by Vimal Kumar Mishra
Answer is no way close to the question.
+1 vote

Hi. As per my knowledge LTE is not Hierarchical architecture instead it has flat architecture that's also one of the reason of giving less delay as compare to other generations.

Simple answer of above query. why we have two different logical entities SGW and PGW just not to mix the signalling data with user data.

3gpp had decided to take care of signaling separately and give the direct tunnel for the User packets.

Like in UMTS/GPRS every packet has to pass through from SGSN and then GGSN even though after signalling info SGSN had no work with user packets. That is the only reason they have separated them to remove unnecessary node from the architecture.

answer Nov 25, 2013 by anonymous
0 votes

sort of load balancing, if SGW is not there performance of PGW node affect due to mobile signaling & handovers.

In roaming condition Visitor SGW communicate with Home PGW of roaming UE.

Usage data of visitor UE generated by SGW in roaming condition.

answer Sep 1, 2017 by Dinesh Yamali
Similar Questions
+1 vote

Hi Friends,
I saw somewhere that MME will also do a DNS query for P-GW. As i know S-Gw will do a query for P-Gw address. Is there any situation that MME will do a query for P-GW address using APN? and after this query how MME will use the P-Gw Address.

and is there any possibility of Doing DNS query by P-Gw to get any other nodes Ips and under what conditions P-Gw will do a DNS query?

0 votes

As per LTE core network architecture, MME shares S11 interface with SGW and SGW has S5/S8 interface with PGW.
I went through Create Session Response message sent by SGW to MME and found PGW-C and PGW-U tunnel ids.
I want to know what MME does with these PGW-C and PGW-U tunnel Ids ?

+1 vote

GTP tunnel id is a 32-bit long value over one IP address combination. Suppose there are two nodes A and B which has IP addresses X and Y respectively. Is it possible when first GTP tunnel id range gets exhausted then both the nodes starts using different IP address combination and again GTP tunnel id range can be re-used ? If it is impossible then what could be the reason ?